How to Cut Big Data Down to Size: Data Driven Marketing Basics
Easy how-to steps to make data driven marketing work for you. Big data statistics, A/B testing variables & direct marketing tips.
April 23, 2013
Big Data. Data Driven Marketing. ROI. Cross-Channel Data. Integrate! Integrate!
Is “Big Data” scaring your pants off? You’re not alone. A Hubspot article on data driven marketing reported that, according to a study reported by eMarketer and conducted by 33Across:
- 91% of survey respondents are concerned about driving ROI from "big data"
- 73% are concerned about integrating cross-channel data
- 70% are concerned about making sense of all the data coming at them
The good news is that data is really your best friend. Sure, you and your team should look at the big picture of your own "big data" and creating a long-term plan, but don’t let that stop you from taking more immediate action. Here are a few easy steps to help you make data-driven marketing work for you.
1. Establish benchmarks.
You can’t know how far you’ve come until you have a good handle on where you started. You can reference general industry benchmarks for KPIs you want to track but ideally you can derive your own specific business benchmarks from your current data. If not, set a plan in place now to start tracking the data you need so you’ll have it going forward.
2. Set SMART marketing goals based on the metrics that matter.
For your direct marketing efforts, you’ll be concentrating on one-to-one communications - so if you are employing specific calls to action, you can track very specific KPIs. But don’t forget to consider lift in social media or website visits and related engagement measures like shares or comments.
3. Test and test again.
Good ol’ A/B testing has never been more valuable. Split your list to test one variable at a time. It’s tempting to test completely different pieces, but do your best to hold yourself to that single variable rule so you can be sure you attribute success to the right thing and can reproduce it. Use commonsense and your best educated guess to choose the variable most likely to impact results. You CAN test everything, but that doesn’t mean you should!
If your list isn’t big enough to produce statistically significant results, consider running the same test across multiple sends so you can look at cumulative results.
4. Put the “drive” into “data-driven.”
After collecting data, you must act on it! This might mean something as simple as reallocating time away from a marketing channel or activity that is less effective into one that shows more results. Or it could be something more complex, like using your data to determine the best ways to segment your contact list for future communications or testing. Our data team at Compu-Mail can be a great resource for help with segmentation and testing. HubSpot also provides some more great ideas on how to take action on your analytics.
This post was written by Karen Renzi, a former Compu-Mail employee.